Although we're trying to prove our antibacterial cream kills bacteria, we need to have some bacteria samples that remain untreated to show that something will grow if they are left alone. Scientists call these extra samples controls. And to overcome inaccurate measuring, replicates are used.

Another beauty of controls and replicates is that simple statistics can be applied. This is particularly important if the result of each sample test was minor. Maybe we would measure the size of the cleared area of each plate and calculate the mean (average) area for each treatment.

How do we know if we've succeeded? 

Major funding for Rough Science was provided by the National Science Foundation. Corporate funding was provided by Subaru.

Copyright © 2002 The Open University and WETA.